Functionalized Pd/ZnO Nanowires for Nanosensorslc/2018-PSS-Pd-ZnO.pdf · 2018-01-15 · enhanced performances. A series of nanosensor devices is fabricated based on single Pd/ZnO

Gas Sensors www.pss-rapid.com

RAPID RESEARCH LETTER

Functionalized Pd/ZnO Nanowires for Nanosensors

Oleg Lupan,* Vasile Postica, Rainer Adelung,* Fr�ed�eric Labat, Ilaria Ciofini,Ulrich Schürmann, Lorenz Kienle, Lee Chow, Bruno Viana, and Thierry Pauport�e

A method for surface doping and functionalization of ZnO nanowires (NWs) withPd (Pd/ZnO) in a one-step process is presented. The main advantage of thismethod is to combine the simultaneous growth, surface doping, and functional-ization of NWs by using electrochemical deposition (ECD) at relatively lowtemperatures (90 �C). Our approach essentially reduces the number of technologi-cal steps of nanomaterial synthesis and final nanodevices fabrication withenhanced performances. A series of nanosensor devices is fabricated based onsingle Pd/ZnO NWs with a radius of about 80nm using a FIB/SEM system. Theinfluence of Pd nominal composition in Pd/ZnO NW on the H2 sensing responseis studied in detail and a corresponding mechanism is proposed. The resultsdemonstrate an ultra-high response and selectivity of the synthesized nano-sensors to hydrogen gas at room temperature. The optimal concentration of PdCl2in the electrolyte to achieve extremely sensitive nanodevices with a gas response(SH2)� 1.3� 104 (at 100ppm H2 concentration) and relatively high rapidity is0.75mM. Theoretical calculations on Pd/ZnO bulk and functionalized surfacefurther validated the experimental hypothesis. Our results demonstrate theimportance of noble metal presence on the surface due to doping andfunctionalization of nanostructures in the fabrication of highly-sensitive andselective gas nanosensors operating at room temperature with reduced powerconsumption.

Nanoscale sizedclusters ofnoblemetals suchasPd,Pt,Au,Ag,Rh,andRu,areknowntoactaseffectivecatalysts[1,2] andplayavital rolein the improvement of the semiconducting oxides sensingproperties and photocatalytic activities.[3–5] A number of methods

Dr. O. Lupan, Prof. R. Adelung, Dr. U. Schürmann, Prof. L. KienleInstitute for Materials ScienceChristian Albrechts University KielKaiserstr. 2, D-24143 Kiel, GermanyE-mail: [email protected]; [email protected]

Dr. O. Lupan, V. PosticaDepartment of Microelectronics and Biomedical EngineeringTechnical University of Moldova168 Stefan cel Mare Av., MD-2004 Chisinau, Republic of Moldova

Dr. O. Lupan, Dr. F. Labat, Dr. I. Ciofini, Dr. B. Viana, Dr. T. Pauport�ePSL Research University, Chimie ParisTech-CNRSInstitut de Recherche de Chimie Paris, UMR824711 rue P. et M. Curie, 75005 Paris, France

Prof. L. ChowDepartment of Physics, University of Central FloridaOrlando, FL 32816-2385, USA

The ORCID identification number(s) for the author(s) of this articlecan be found under https://doi.org/10.1002/pssr.201700321.

DOI: 10.1002/pssr.201700321

Phys. Status Solidi RRL 2018, 12, 1700321 © 21700321 (1 of 9)

has been proposed to incorporate thesemetals into semiconducting oxides micro-andnanostructures inorder to improve theirUV- and gas-sensing properties in the mostefficient way. The most well-documentedmethods are adsorption, doping,[6,7] surfacefunctionalization (decorating, hybridiza-tion, loading, impregnating),[3,8] and com-posing.[9] Among all noble metals, Pd-doping/incorporation has been demon-strated to show the highest efficiency inthe detection of H2 gas.

[2,5,10] Such radicallyimproved performances can be attributed toboth chemical and electronic sensitiza-tions.[5] The presence of Pd nanoparticleson ZnO greatly improves the room temper-ature catalytic activity due to the high H2

solubility in Pd which gives higher concen-tration of clusters (catalytic centers) andlowers the saturation rate of response andrecovery processes.[11]

The detection mechanism of sensorsbased on NWs in most cases is based onthe modulation of the conduction channel/electron depleted region (EDR).[3,6] In thecase of surface functionalization, the

Schottky barriers are formed at the Pd/semiconducting oxideinterface due to higher work function of Pd compared toZnO,[3,12] leading to more narrowed conduction channel(extension of EDR). After exposure to H2 gas, the formationof PdHx phases with lower work function can take place,[2,12]

which lower the height of the Schottky barrier and expand theconduction channel width (suppression of EDR).[6,12] This typeof mechanism is related to electronic sensitization.[5] Anotherimportant sensing mechanism is related to chemical sensitiza-tion and is based on dissociation of H2 molecules into Hatoms,[2,13] which interact with adsorbed oxygen species onto thesurface of semiconducting oxide nanostructures. It is not wellunderstood yet which mechanism is more dominant underdifferent operating conditions.[3,14] However, in the case ofsurface functionalization with noble metal nanoparticles (NPs)due to the dependence of gas response on size and homogeneityof NPs, density, and space distribution, some issues still need tobe resolved in order to improve long term stability andrepeatability of sensors.[3,14] Thus, different methods for efficientcontrol of the NPs deposition are still investigated.[14]

In this study we demonstrate the outstanding increase inhydrogen gas sensing properties of a single Pd/ZnO NW basednanosensor after surface doping and functionalized with PdNPs. Both surface doping and functionalization are performed

017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

https://doi.org/10.1002/pssr.201700321

http://www.pss-rapid.com

www.advancedsciencenews.com www.pss-rapid.com

during the NWs growth, which is the main advantage of thisproposed method. This effectively reduces the fabrication stepsof the final device and excludes many multistep technologies ofsemiconducting oxides manufacturing. Pd-doped ZnO bulk andfunctionalized surfaces were also modeled using DFT to obtaininsights into the experimental findings. Devices were fabricatedvia focused ion beam (FIB)/SEM system and showed the ultra-sensitive and highly selective response at room temperature withthe lowest detection limit (LDL) in the range of sub-ppm(0.015 ppm or 15 ppb). The effect of Pd content was studied indetail, and the respective sensing mechanism was proposedbased on theoretical and experimental data. Presented results areof high scientific interest, especially in the field of nanodevicefabrication and for highly sensitive nanosensors.[2,3,15,16] In thisstudy the influence of junctions between ZnO NWs on gasresponse (important in the case of networks) is excluded andonly the influence of functionalization is investigated. Thenovelty of this work also lies in getting further insights into thegas sensing mechanism of individual hybrid Pd/ZnO NWs,which is very important from a fundamental point of view.

Experimental Section: The ZnO NWs were grown usingelectrochemical deposition in a three-electrode electrochemicalcell at 90 �C as was reported in previous works.[6,17,18,19]

Synthesis, surface doping, and functionalization with Pd NPsof ZnO NWs were achieved in one step-process by adding PdCl2solution (Alfa Aesar) in the electrolyte solution with aconcentration range of 0.25–1.50mM. F-doped SnO2 (FTO)polycrystalline films (sheet resistance 10Ω sq) were used assubstrates and as working electrode (WE) for the electrodeposi-tion of Pd/ZnO NW arrays.[17,18] The procedure of substratescleaning was reported in previous works.[17,18] During thedeposition process, the angular speed of WE was set atω¼ 300 rotationmin�1 (rpm). Total deposition time was 7000–9000 s.[17] The Pd/ZnO NW arrays after deposition werethermally annealed at 250 �C for 12 h. Applied voltage fordeposition was determined from voltammetry cycles (notshown)[17] and are presented in Table S1, Supporting Informa-tion. Figure S1, Supporting Information, shows growth curvesfor samples with different concentrations of PdCl2 in theelectrolyte at respective constant applied potential. Taking intoaccount values of current density reported for pure ZnONWs,[18]

by adding PdCl2 in the electrolyte a dramatic increase of thedeposition cathodic current density is obtained, which was alsoreported for ZnO:Ag NWs.[17] Thus, we believe that Pd on thedeposited structures serves as a good electrocatalyst for theelectrochemical reactions. The experimental details on materialcharacterization are presented in Supporting Information (TextS2). Computational details about DFT modeling and choice ofthe functionalized surface models are presented in SupportingInformation (Text S3 and S4, respectively).

Morphological and Structural Properties of Pd/ZnO NW Arrays:Figure 1 shows SEM images at different magnifications ofPd/ZnO NW arrays prepared from various concentrations ofPdCl2 in electrolyte: (a–c) 0.25mM; (d–f) 0.50mM; (g–i) 0.75mM;(j–l) 1.00mM; (m–o) 1.25mM; (p-r) 1.50mM. They are displayedfrom lowmagnification in the first column to highmagnificationin the third column. The details on pristine ZnO NWarrays andmore informations about the influence of deposition parameterson the morphology are presented in our previous works.[17–19]

Phys. Status Solidi RRL 2018, 12, 1700321 1700321 (

The mechanism of ZnO NWs electrodeposition was alreadydiscussed in a previous work.[18] The ZnO NWs have diametersin the range of D¼ 100–200 nm and lengths of �2.8–3.1mm.[17]

By adding PdCl2 into the electrolyte, changes in aspect ratio andmorphology of the NWs were observed (see Figure 1). Thehexagonal shape of Pd modified ZnO NWs gradually disappearsas the PdCl2 concentration in the electrolyte increases toward thehighest content. This tendency was also observed for Cd/ZnONW arrays by ECD.[18] The diameter of the NWs is decreased toabout 100–150 nm when the PdCl2 concentration is increased to0.75mM leading to more dense networks, while for furtherincrease of the PdCl2 concentration up to 1.5mM, the diameter ofthe NWs considerably increases up to �500 nm (see Figure1p–r). The enhanced density of ZnO NWs was also observed inthe case of Ag-doped ZnO NWs.[17] It can be attributed to theelectrocatalytic properties of PdCl2 which favors the hydroxideion generation at the beginning of the electrochemicalgrowth.[17] More representative SEM images of Pd modifiedZnO NW arrays at low magnification to show overall view ofnetworks and at much higher magnifications are presented inFigure S2, Supporting Information, for 0.25mM, Figure S3,Supporting Information, for 0.50mM, Figure S4, SupportingInformation, for 0.75mM, and Figure S5, Supporting Informa-tion, for 1.0mM PdCl2. Such rough surface and further increasein diameter (D) of Pd modified ZnO NW (for concentrations ofPdCl2 higher than 0.75mM), that is laddered side surfaces andquite tapered tips of NWs (see Figure 1) can be attributed to thedisturbance of the crystal growth in solution, where formation ofnative defects such as vacancies migrate to the surface and formpits, which was also observed for Cl- and Sb-doped ZnO NWs.[20]

Figure 2a shows the effect of the Pd content in Pd/ZnO NWarrays on the crystallinity. The reflections marked with redasterisk (�) are assigned to tetragonal SnO2 from FTO substrateaccording to JCPDS 01-088-0287 card.[17] Others reflections areassigned to hexagonal ZnO crystalline phase according to JCPDS00-036-1451 card. Reflections attributed to Pd, PdO or impuritieswere not observed, indicating that addition of PdCl2 in theelectrolyte does not change essentially the crystal structure ofPd/ZnO NWs.[17] However, the reflections which are commonlyobserved in XRD patterns for Pd modified metal oxidenanostructures have the same values as ZnO main reflectionsand overlap it, thus it is difficult to distinguish Pd or PdO fromZnO by XRD.[21] Higher intensity of reflection corresponding to(002) plane of the samples shows that the growth is along the c-axis normal to substrate as unambiguously demonstrated byTEM (see Figure S6, Supporting Information).[17,22] The changein intensity of XRD reflection corresponding to (002) plane canbe observed with increasing concentration of PdCl2 inelectrolyte, and this can be attributed to modifications incrystallinity of Pd/ZnO NWs.

No significant shift of reflections from XRD diffractograms byincreasing PdCl2 concentration in the electrolyte was observed.This suggests no detectable change of crystal lattice as a result ofthe similar ionic radii of Pd and Zn (r(Pd2þ)¼ 0.078 nm while r(Zn2þ)¼ 0.074 nm).[23] Thus, more likely Pd is not incorporatedin the lattice of Pd-modified ZnO NWs during electrodeposition,although it can diffuse during the annealing process. To the bestof our knowledge, reports with fair evidence of Pd doping in ZnONWs are yet to be seen. This is probably due to low solubility of

© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim2 of 9)

http://www.advancedsciencenews.com


Figure 1. SEM images of Pd/ZnO NW arrays at different magnifications (from lower magnification in the first column to higher magnifications in thethird column) and with different concentrations of PdCl2 in the electrolyte (in mM): (a–c) 0.25; (d–f) 0.50; (g–i) 0.75; (j–l) 1; (m–o) 1.25; (p–r) 1.50.


Pd in ZnO.[20] At the same time we see a bandgap shift(Figure 2b,c).

PdmodifiedZnONWs (with 0.75μMPd)were examinedusingTEM. Figure S6a, Supporting Information, shows the expectedneedle-like structure. The high resolution TEM (HRTEM)micrograph (the average background subtraction filter ABSFfiltered) and the resulting fast Fourier transform (FFT) (seeFigure S6b, Supporting Information) indicate that a needle-like


crystal grows along the c-direction. Figure S6c, SupportingInformation, displays the selectedareaelectrondiffraction (SAED)pattern of a relatively larger and isolated ZnO needle (zone[2 -1 -1 0]). The measured d-values convincingly agree with valuesfrom literature for ZnO.[24]

High angle annular dark field (HAADF)-STEM imagesreveal the presence of particles on the Pd modified ZnO NWs(0.75mM of PdCl2) with a size of about 20 nm. The density of




Figure 2. (a) XRD diffractograms; and (b) optical transmittance spectra of Pd/ZnONW arrays. (c) Plot of (αhν)2 versus photon energy (hν). Inset: energybandgap (Eg) versus concentration of PdCl2 in electrolyte. (d) Room-temperature photoluminescence spectra of Pd/ZnO NW arrays.


these particles on the surface differs for different samples (seeFigure 3a and e). Energy-dispersive X-ray spectroscopy (EDX)elemental maps show that these nanoparticles contain the Pdelement (see Figure 3b–d). The oxidation state of Pd will bediscussed based on X-ray photoelectron spectroscopy (XPS) databelow, and we will show that these NPs are mainly formed bymetallic Pd with low content of PdO.

The formation of Pd NPs on ZnO NWs surface can beexplained based on electrodeposition from electrolyte solutionwith added PdCl2 (Pd

2þðaqÞ þ 2e� ! PdðsÞ), which occurs withhigh current efficiency.[5] However, the formation of films isstopped due to formation of Schottky barriers at Pd/ZnO NWinterface, which blocks further deposition of Pd resulting information of NPs.[25] This was also observed for other types ofsemiconductors.[25]

Optical and Chemical Properties of Pd/ZnO NW Arrays:Figure 2b shows the room temperature UV-visible transmissionspectra of Pd/ZnO NWs with different PdCl2 concentrations inthe electrolyte. A typical ZnO adsorption edge in the UV regioncan be observed for all samples. Similar to Ag/ZnO NWs andCu/ZnO NWs, the decrease in the visible and near-infrared(NIR) range can be observed.[17,25] For samples synthesized with0.25mMPdCl2 in the electrolyte the transmittance is higher than75% in the visible region and close to 90% in the near-infrared


region. By increasing the concentration of PdCl2 in theelectrolyte up to 1.25mM the transmittance is decreased to40% in the visible region and close to 60% in the NIR region.Figure 2c shows that the energy bandgap (Eg) shifts with higherPdCl2 concentration in the electrolyte, thus making thesematerials quite interesting for optical applications.[17,18,19,25] Theinset of Figure 2c presents the estimated Eg versus concentrationof PdCl2 in the electrolyte. Because the incorporation of Pd intoZnOmatrix was excluded, the slight decrease in bandgap energycan be explained by a change in the carrier concentration due toaddition of Pd NPs.[26] Also it was demonstrated that the additionof PdCl2 in the reaction system can result in a reduction ofoxygen vacancy concentration.[26]

A considerable decrease in the intensity of PL emission atroom temperature with an increase of the PdCl2 concentration inthe electrolyte can be observed in Figure 2d. However, for allsamples no significant emission in the visible range due todefects is observed, demonstrating that Pd/ZnONWs are of highquality, cf. supporting information in Figure S6a.[17] Beside thedecrease in the intensity of the near-band edge (NBE) emissionpeak, a slight shift toward higher wavelengths occurs with theincrease of the Pd content.

Since TEM measurements showed the presence of Pd NPs inPd modified ZnO NWs, the surface of the samples was further




Figure 3. (a) HAADF-STEM image of ZnO crystals with few palladium nanoparticles. EDX elemental mapping showing: (b) Zn (L-line); (c) O (K-line);and (d) Pd (L-line) maps of the emphasized region in (a). (e) HAADF-STEM image of ZnO crystals with a higher amount of palladium nanoparticles froman additional sample. Higher resolution XPS spectrum of: (f) O-1s and (g) Pd-3d5/2 obtained from Pdmodified ZnONW arrays. SIMSmeasurements of:(h) uncoated and (i) photoresist-coated Pd modified ZnO NW arrays (0.75 μM Pd) grown on FTO substrate.


analyzed by XPS in order to determine the oxidation state ofPd-NPs in more detail. The XPS sample was first partially spin-coated with S1813 photoresist at 300 RPM for 60 s. After spin-coating, the sample was annealed at 115 �C for 60 s. In XPSsurvey spectra of uncoated sample grown by adding 0.75mM ofPdCl2 in the electrolyte without pre-cleaning procedure(Figure S7, Supporting Information), Zn, O, Pd, C elementswere detected. A relatively high concentration of adventitiouscarbon (�35%) was detected.[22] Inset in Figure S6, SupportingInformation shows the XPS spectra of the Zn-2p2/3 core levelregions of ZnO:Pd NWs at 1022.2 eV. The asymmetric O-1s peakfrom Figure 3f was deconvoluted by two subspectral componentscorresponding to ZnO (530.3 eV) and defective ZnOx and/or Zn-OH species (532.6 eV).[22] The line intensity of Pd-3d5/2 core levelemission peaks of the Pd modified ZnO NWs is presented inFigure 3g. A detailed peak analysis suggests a superposition oftwo peaks located at 334.9 and 336.1 eV which are attributed toPd and PdO, respectively (see Figure 3g).[27] The oxide formationof Pd can be the result of the interaction with water vapor inambient air or interaction with oxygen from growth solutionduring the ECD deposition.[17,28]

The chemical composition of Pd modified ZnO NW arrays(0.75mM of PdCl2) grown on FTO substrate was furtherinvestigated using SIMS measurements. Figure 3h shows theresults of uncoated arrays, that is Zn, Sn, Pd, Si, and C signalsversus sputtering time. Sn signal originated from FTO substrateappears very early in the measurement and do not follow the rateof Zn signal, which means that the ZnO NW arrays are not verydense, so the ion beam can sputter the bottom substrate(FTO).[6,16] At sputtering time of 1000 s, the Si signal starts toincrease, which means the FTO film on the glass, has been


removed due to sputtering, so the Si signal from the glasssubstrate appears. The C signal is originating from surfacecontamination of top ZnO NW arrays, which is confirmed by avery sharp decrease at the beginning of sputtering.[16] In the caseof photoresist coated Pd/ZnO NW arrays (see Figure 3i) the Snsignal does not appear until 2000 s and then rises very sharply. Itmeans that the photoresist prevents the ion beam frompenetrating into the aligned ZnO NWs. The C signal with highintensity originates from photoresist, and it is more or lessuniform until one penetrates into the glass. As in the case ofuncoated samples, the Pd signal tracks very close Zn signal. Theexperiment suggests that the Pd content is about 0.15–0.20 at.%in this sample.[6,16] Also, the Si signal in the case of photoresistcoated samples starts around 3000 seconds. This occurs whenthe photoresist and FTO substrate both are sputtered away andthe glass signal (Si) starts to appear.

Modulation of Structural and Electronic Properties UponDoping � Theoretical Insights: From the above discussions, thePd-doping in bulk of ZnO nanowires was not demonstrated.However, traces of surface doping and functionalization of Pd/ZnO NWs, can be clearly seen. Consequently, to get a betteroverview of possible Pd/ZnO structures, we present dataobtained on models of Pd-doped bulk ZnO wurtzite andadsorption of Pd clusters of different sizes on the undoped anddoped ZnO (1 0 -1 0) surface, as an example of surfacefunctionalization. As already done for Cd,[16,18] Cu,[29] andAg,[8,17] calculations have been performed on: (i) pure ZnO, (ii)Pd-doped ZnO, at three different amounts (0.78, 1.85, and6.25 at.%) and Pdi only at the 1.85 at.% amount, as an example ofthe effect of the Pd amount on structures and electronicproperties. Obtained data are collected in Table S2, Supporting





Information, and unit cells are shown in Figure S8, SupportingInformation.

From the computed shortest metal to O distances (d1 to d6 inTable S2, Supporting Information), when substituting, it is clearthat the coordination of the metal goes from tetrahedral in pureZnO with significantly different values for d1 and d5 (seeFigure S8e, Supporting Information), to an almost regulartriangular bipyramid when considering substitution at 6.25 at%,with similar d1 and d5 values. Consequently, an increase of in thea and b lattice parameters and a decrease of in c are obtained inthe doped material when compared with pure ZnO. Uponinsertion however, a distorted octahedron can be evidenced, withthree long (�2.7 A) and three shorter distances (�2.3 A). In thatcase, all three lattice vectors are slightly increased, due to thelarger Pd radius compared to that of Zn, evidencing that Pdincorporation in the experimental samples is probably mostlydue to both substitution and adsorption. These geometricaldistortions can all be related to a Jahn-Teller effect. From the dataof Figure S9 and Table S2, Supporting Information, it is clearthat the ZnO band gap decreases upon Pd contents, by at least1.2 eV in agreement with experimental data (therefore themodeling shows the Pd doping). For substitution, Pd contributesmainly at the top of the valence band (VB) and the bottom of theconduction band (CB), establishing bands with low dispersionand an almost atomic-like character. However, for concentra-tions of 6.25 at.% a small overlap with O2p states can beevidenced, indicating a larger hybridization of Pd4d and O2p

orbitals. When substituting, a shallow level can be found in theZnO band gap indicating p-type ZnO formation, as alreadyobtained in the case of Ag doping.[17] Similar conclusions can bedrawn for insertion, but the lack of shallow level at the bottom ofthe CB indicates that no p-type ZnO is obtained.

Adsorption of small Pdn clusters (n ranging from 1 to 9) on aZnO (1 0 -1 0) slab[17,30] has also been considered (see Figure 4).

Figure 4. Side and top views of the unit cells of the pure ZnO (1 0 -1 0) surfacgrey and blue balls represent O, Zn, and Pd atoms, respectively. Unit cell a


Comments on the functionalized surfacemodels can be found inthe Supporting Information (Text S3).

From the data collected in Table S3, Supporting Information,it is clear that the larger the Pdn cluster is, the larger is theadsorption energy and the smaller the electronic band gap.Larger clusters show a larger spread over the ZnO surface,leading to larger adsorption energies due to the increasingnumber of Pd-O and Pd-Zn bonds between the adsorbate and thesubstrate. For Pd1 and Pd2 clusters, Pd preferentially adsorbs inbridge sites that are in-between Pd and O atoms. For largerclusters, hollow sites are populated, leading to an hexagonalclose packed (hcp) motif on top of the ZnO substrate. Althoughthe obtained band gap values for all adsorbed systems differsignificantly from those reported experimentally in Figure 2c, itshould be noted that: (i) the trend of decreasing Eg upon increaseof the PdCl2 concentration in the electrolyte is qualitativelyreproduced; (ii) the computed band gap of the pure ZnO (1 0 -1 0)surface of 3.89 eV nicely reproduces that of the pure ZnO NWobtained by experiments,[18] thus validating our model.Increasing concentration of Pd leading to atomic substitutionwithin the ZnO slab (system Pd9/ZnO) has also been considered,but only a slight variation of computed band gaps could beevidenced (<10%).

From Figure S10, Supporting Information, as already foundfor the bulk system, weakly-interacting Pd4d levels contributemainly to the band gap decrease observed for the functionalizedsurface. However, shallow levels can only be evidenced at thebottom of the CB for clusters larger than Pd3, outlining the roleof the NP size in determining the electronic structure of thesesurfaces, and hence the possibility of strongly influencing theirreactivity, by tuning the Pd cluster sizes.

High Performance Nanosensor Based on Individual Pd/ZnONWs: Based on single Pd/ZnO NW with different Pd contents,NWswith diameter (D) of approximately 160nmand length about

e with 8 layers, and of Pdn/ZnO (1 0 -1 0), with n ranging from 1 to 9. Red,s solid yellow line is shown.





1.4 μmwere fabricated (i.e., practically the same aspect ratio), andseveraldeviceswere testedat100ppmH2at roomtemperatureandrelative humidity (RH) of 30%RH. It is important tomaintain thesame aspect ratio because the diameter of individual nano-structures has a larger influence on gas sensing properties.[6,16]

Figure 1 demonstrates that the PdCl2 content in electrolytesolution influences the aspect ratio ofPd/ZnONWs.Thus, if itwasthe case of sensors based on arrays of Pd/ZnO NWs, the aspectratio should have been taken into account.

Results on hydrogen response of single ZnO NWnanosensors(diameter of �160nm) versus PdCl2 concentration and gasresponse of nanosensors to different types of reducing gases arepresented in Figure 5a,b. Inset of Figure 5b shows a typical SEMimage of nanosensors based on a single Pd/ZnO NW withD� 160nmand grown in the electrolytewith 0.75mMofPdCl2. InFigure S11, Supporting Information, a typical current-voltagecharacteristic at room temperature of the nanosensor device (Pd/ZnO NW grown in the electrolyte with 0.75mM of PdCl2) ispresented. Formation of double Schottky contacts is due to the Ptcontacts, thathashigherworkfunctionthanthatofPtcontactwhencompared to ZnO.[31] All fabricated devices in this study showedthe same behavior of current-voltage characteristics (not shown).

As can be observed from Figure 5a, the hydrogen response(IH2/Iair) is promoted and strongly enhanced from�1.5 to about13100 (about four orders of magnitude increase!) by increasingPdCl2 concentration in the electrolyte from 0.25 to 0.75mM. Afurther increase in PdCl2 concentration to 1.5mM resulted in adecrease down to�9.2. Thus, the optimal PdCl2 concentration inthe electrolyte for growth of Pd/ZnO NWs is determined to be0.75mM (see Figure 5a). The response value is much higher thanfor single pristine ZnO NW, Ag/ZnO,[8,17] and Cd/ZnO,[16] NWor other single and networked semiconducting nanostructures(see Table S4, Supporting Information). Next the selectivity of a

Figure 5. (a) Hydrogen response of a single ZnO NW nanosensor with a dresponse of nanosensors to different types of reducing gases. The inset is ahydrogen gas response to single and multiple pulses are presented, respec


nanosensor based on Pd/ZnO NW (0.75mM of PdCl2) wasinvestigated by exposure to 100 ppm of CO2, CH4, ethanol andacetone vapor (see Figure 5b). To evaluate the selectivity of H2

gas nanosensor the ratio of the gas responses were used, i.e. theratio of hydrogen gas response (SH2) and of gas response toother gases. The results of SH2/SCO� 2848, SH2/SCH4� 2339,SH2/SEtOH� 1065, and SH2/SAcetone� 2079 show excellentselectivity to H2 gas.

The dynamic response of the nanosensors is presented inFigure 5c for a device based on a Pd/ZnO NW grown in theelectrolyte solution with 0.75mM of PdCl2, and in Figure S12,Supporting Information, for devices with different concentra-tions of PdCl2, respectively. Up to 0.75mM the nanosensorresponse is rapidly approaching the saturation value (seeFigure 5c and Figure S12a,b, Supporting Information), whilethe response of devices based on a Pd/ZnO NW grown in theelectrolyte solution with concentration of PdCl2 higher than0.75mM is still increasing even after exposure to H2 gas for 30 s(see Figure S12c,d, Supporting Information). The response time(tr) and recovery time (td) were calculated and generalized inFigure S13a and Table S5, Supporting Information, which clearlyshow that tr and td values are decreasing up to 0.75mM of PdCl2in the electrolyte and that these values are increasing with afurther increase of the PdCl2 concentration. This can beexplained based on an enhanced rate of catalytic activity ofPd/ZnO NW by increasing Pd content.[10,12,14] However athigher content of Pd in a Pd/ZnO NW, the catalytic activity ismuch pronounced which gives higher concentration of catalyticcenters and lowers the saturation rate of the response andrecovery process.[10,12,14]

The lowest detection limit (LDL) for H2 detection of ananosensor based on Pd/ZnO NW (0.75mM of PdCl2) wasapproximated by linear fitting from Figure S13b, Supporting

iameter of �160 nm versus concentration of PdCl2 in electrolyte. (b) GasSEM image of the nanosensor (the scale bar is 200 nm). (c,d) Dynamictively.





Information, with response criteria Igas/Iair> 1.2 and showedsub-ppm value beyond 0.015 ppm. This value is lower than thatfor previously reported sensors on Fe-doped ZnO[22] or Zn-dopedCuO and Cu2O nanostructured films,[6,32] Pd coated Si NWs,[33]

single TeO2 NW.[34] This finding demonstrates that ouroptimized Pd/ZnO NW sensor possesses one of the lowestdetection limits among oxide-based hydrogen sensors.[10,22,33,34]

Also, the detection limit is much lower than that reported formost of the commercial hydrogen gas sensors for industrial,environmental, and military monitoring.[33] This outstandingperformance could be based on the very high surface-to-volumeratio of the NWs and the high catalytic activity of Pd.[7,13,15,33] Inconclusion, the optimal content of PdCl2 in the electrolyte forgrowth of Pd/ZnO NW with ultra-sensitive and highly selectivedetection of H2 gas is 0.75mM. Figure 5d shows the repeatabilityand stability of the fabricated nanosensor based on a Pd/ZnONW (0.75mM of PdCl2 and D� 160 nm) with a deviation in thegas response lower than 5%. The gas sensing mechanism isdiscussed in Supporting Information (Text S1) and was alsodiscussed in detail in our previous work.[35] The high selectivityto H2 gas of Pd-modified ZnO NWs can be explained based onlow hydrogen-binding energy of Pd, as well as its low energybarriers for hydrogen dissociation with further formation ofPdHx with different properties.[36]

Another important factor is the formation of the Schottkycontacts (at the Pt/ZnO NW interface) which is known to behighly selective to hydrogen gas.[37] After catalytic chemicaladsorption of H2 gas molecules on Pt with further diffusionthrough Pt layer, a dipole layer is formed at the Pt/ZnOinterface.[37] This reduces the Schottky barrier height andinduces a considerable change in the device current.[37] Moredetails on this mechanism are presented in Refs. [37].

In summary, Pd/ZnO NW arrays were synthesized onto FTOsubstrate by electrodeposition at 90 �C using a method forsurface doping and functionalization in a one-step process. Theadvantage of this procedure is to combine growth, surfacedoping, and functionalization of Pd/ZnO NWs by using ECD,and to reduce essentially the number of technological steps fornanomaterial and final nanodevices fabrication. Differenttechniques of material characterization, such as XRD, Raman,XPS, and TEM have demonstrated the high crystallinity of thePd/ZnO NWs synthesized over a wide range of PdCl2concentrations in the electrolyte (0.25 to 1.50mM). SingleNWs of Pd/ZnO with different concentrations of Pd wereintegrated into nanosensors using a FIB/SEM system. The gassensing properties were investigated in detail. The influence ofthe PdCl2 concentration was studied and revealed a high gasresponse for Pd/ZnO NW (D� 160 nm) grown in the electrolytewith 0.75mM of PdCl2. An ultra-high response or a giantresponse of Igas/Iair� 13100 were obtained to 100 ppm of H2 gasat room temperature and 30% RH. Excellent selectivity wasdemonstrated by testing with other types of reducing gases, anddemonstrated negligible response compared to those of H2 gas.The excellent H2 gas sensing properties of the Pd/ZnO NWhavebeen explained based on improved catalytic properties due to PdNPs surface functionalization. DFT calculations performed onPd-doped ZnO bulk and functionalized surface models allowedto support and further validate the model proposed based on theexperimental findings.


Supporting InformationSupporting Information is available from the Wiley Online Library or fromthe authors.

AcknowledgementsDr. Lupan acknowledges the Alexander von Humboldt Foundation for theresearch fellowship for experienced researchers 3-3MOL/1148833 STP atthe Institute for Materials Science, University of Kiel, Germany. Thisresearch was funded partially by the German Research Foundation (DFG)under the scheme SFB 1261 and FOR2093. This work was partiallysupported by the STCU within the Grant 6229.

Conflict of InterestThe authors declare no conflict of interest.

Keywordsgas sensors, hydrogen, nanowires, palladium, ZnO

Received: September 15, 2017Revised: November 4, 2017

Published online: November 20, 2017

[1] (a) T. K. Sau, A. L. Rogach, Adv. Mater. 2010, 22, 1781; T. K. Sau, A. L.Rogach, F. Jäckel, T. A. Klar, J. Feldmann, Adv. Mater. 2010, 22, 1805;(b) M. V. Seregina, L. M. Bronstein, O. A. Platonova, D. M.Chernyshov, P. M. Valetsky, J. Hartmann, E. Wenz, M. Antonietti,Chem. Mater. 1997, 9, 923; (c) W. Shenton, S. A. Davis, S. Mann, Adv.Mater. 1999, 11, 449.

[2] F. Favier, E. C. Walter, M. P. Zach, T. Benter, R. M. Penner, Science2001, 293, 2227.

[3] A. Kolmakov, D. O. Klenov, Y. Lilach, S. Stemmer, M. Moskovits,Nano Lett. 2005, 5, 667.

[4] S. Sakthivel, M. V. Shankar, M. Palanichamy, B. Arabindoo,D. W. Bahnemann, V. Murugesan, Water Res. 2004, 38, 3001.

[5] (a) N. Yamazoe, Sens. Actuators B 1991, 5, 7; (b) F. Yang,D. K. Taggart, R. M. Penner, Nano Lett. 2009, 9, 2177.

[6] O. Lupan, V. Cretu, V. Postica, O. Polonskyi, N. Ababii, F. Schütt,V. Kaidas, F. Faupel, R. Adelung, Sens. Actuators B 2016, 230, 832.

[7] M. Hübner, N. Barsan, U. Weimar, Sens. Actuators B 2012, 171-172,172.

[8] C. Liu, Q. Kuang, Z. Xie, L. Zheng, CrystEngComm 2015, 17, 6308.[9] X. Liu, N. Chen, B. Han, X. Xiao, G. Chen, I. Djerdj, Y. Wang,

Nanoscale 2015, 7, 14872.[10] E. SSennik, O. Alev, Z. Z. €Oztürk, Sens. Actuators B 2016, 229, 692.[11] M. G. Chung, D.-H. Kim, D. K. Seo, T. Kim, H. U. Im, H. M. Lee,

J.-B. Yoo, S.-H. Hong, T. J. Kang, Y. H. Kim, Sens. Actuators B 2012,169, 387.

[12] N. Y. Chan, M. Zhao, J. Huang, K. Au, M. H. Wong, H. M. Yao, W. Lu,Y. Chen, C. W. Ong, H. L. W. Chan, J. Dai, Adv. Mater. 2014, 26, 5962.

[13] Y. C. Lee, H. Huang, O. K. Tan, M. S. Tse, Sens. Actuators B 2008,132, 239.

[14] C.-M. Chang, M.-H. Hon, I.-C. Leu, ACS Appl. Mater. Interfaces 2013,5, 135.

[15] Y. Hu, J. Zhou, P.-H. Yeh, Z. Li, T.-Y. Wei, Z. L. Wang, Adv. Mater.2010, 22, 3327.





[16] (a) O. Lupan, L. Chow, T. Pauport�e, L. K. Ono, B. Roldan, Cuenya,G. Chai, Sens. Actuators B 2012, 173, 772; (b) O. Lupan, V. Postica,J. Gröttrup, A. K. Mishra, N. H. de Leeuw, R. Adelung, Sens. ActuatorsB 2017, 245, 448; (c) F. Schütt, V. Postica, R. Adelung, O. Lupan, ACSAppl. Mater. Interfaces 2017, 9, 23107.

[17] T. Pauport�e, O. Lupan, J. Zhang, T. Tugsuz, I. Ciofini, F. Labat,B. Viana, ACS Appl. Mater. Interfaces 2015, 7, 11871.

[18] (a) O. Lupan, T. Pauport�e, T. Le Bahers, I. Ciofini, B. Viana, J. Phys.Chem. C 2011, 115, 14548; (b) T. Pauport�e, D. Lincot, J. Electrochem.Soc. 2001, 148, C310; (c) A. Goux, T. Pauport�e, J. Chivot, D. Lincot,Electrochim. Acta 2005, 50, 2239; (d) T. Pauport�e, G. Bataille,L. Joulaud, F. J. Vermersch, J. Phys. Chem. C 2010, 114, 194.

[19] (a) O. Lupan, B. Viana, T. Pauport�e, M. Dhaouadi, F. Pell�e, L. Devys,T. Gacoin, J. Phys. Chem. C 2013, 117, 26768; (b) O. Lupan, T.Pauport�e, V. V. Ursaki, I. M. Tiginyanu, Optic. Mater. 2011, 33, 914;(c) O. Lupan, T. Pauport�e, I. M. Tiginyanu, V. V. Ursaki, H. Heinrich,L. Chow, Mater. Sci. Eng. B 2011, 176, 1277; (d) O. Lupan, T.Pauport�e, B. Viana, P. Aschehoug, M. Ahmadi, B. R. Cuenya, Y.Rudzevich, Y. Lin, L. Chow, Appl. Surf. Sci. 2013, 282, 782.

[20] (a) F. Wang, J.-H. Seo, Z. Li, A. V. Kvit, Z. Ma, X. Wang, ACS Appl.Mater. Interfaces 2014, 6, 1288; (b) J. K. Liang, H. L. Su, C. L. Kuo,S. P. Kao, J. W. Cui, Y. C. Wu, J. C. A. Huang, Electrochim. Acta 2014,125, 124; (c) O. Lupan, T. Pauport�e, L. Chow, B. Viana, F. Pell�e,B. Roldan Cuenya, L. K. Ono, H. Heinrich, Appl. Surf. Sci. 2010,256, 1895.

[21] Q. Li, Y. W. Li, P. Wu, R. Xie, J. K. Shang, Adv. Mater. 2008, 20, 3717;Q. Li, R. Xie, E. A. Mintz, J. K. Shang, J. Am. Ceram. Soc. 2007,90, 3863.

[22] I. Hölken, G. Neubüser, V. Postica, L. Bumke, O. Lupan, M. Baum,Y. K. Mishra, L. Kienle, R. Adelung, ACS Appl. Mater. Interfaces 2016,8, 20491.

[23] R. Shannon, Acta Crystallogr. A 1976, 32, 751.[24] H. Sowa, H. Ahsbahs, J. Appl. Crystallogr. 2006, 39, 169.


[25] (a) H. Hasegawa, T. Sato, Electrochim. Acta 2005, 50, 3015;(b) O. Lupan, T. Pauport�e, B. Viana, P. Aschehoug, Electrochim.Acta 2011, 56, 10543.

[26] Y. Chang, J. Xu, Y. Zhang, S. Ma, L. Xin, L. Zhu, C. Xu, J. Phys. Chem. C2009, 113, 18761.

[27] G. Ketteler, D. F. Ogletree, H. Bluhm, H. Liu, E. L. D. Hebenstreit,M. Salmeron, J. Am. Chem. Soc. 2005, 127, 18269.

[28] M. Moroseac, T. Skála, K. Veltruská, V. Matolın, I. Matolınová, Surf.Sci. 2004, 566-568, Part 2, 1118.

[29] O. Lupan, T. Pauport�e, T. Le Bahers, B. Viana, I. Ciofini, Adv. Funct.Mater. 2011, 21, 3564.

[30] F. Labat, I. Ciofini, C. Adamo, J. Chem. Phys. 2009, 131, 044708.[31] Z. L. Wang, J. Song, Science 2006, 312, 242.[32] V. Cretu, V. Postica, A. K. Mishra, M. Hoppe, I. Tiginyanu,

Y. K. Mishra, L. Chow, N. H. de Leeuw, R. Adelung, O. Lupan, J.Mater. Chem. A 2016, 4, 6527.

[33] J.-S. Noh, H. Kim, B. S. Kim, E. Lee, H. H. Cho, W. Lee, J. Mater.Chem. 2011, 21, 15935.

[34] C. Yan, B. Huy Le, D. J. Kang, J. Mater. Chem. A 2014, 2, 5394.[35] O. Lupan, V. Postica, F. Labat, I. Ciofini, T. Pauport�e, R. Adelung,

Sens. Actuators B 2018, 254, 1259.[36] (a) T.-R. Rashid, D.-T. Phan, G.-S. Chung, Sens. Actuators B 2013, 185,

777; (b) X. Q. Zeng, M. L. Latimer, Z. L. Xiao, S. Panuganti, U. Welp,W. K. Kwok, T. Xu, Nano Lett. 2011, 11, 262; (c) V. Postica, F. Schütt,R. Adelung, O. Lupan, Adv. Mater. Interfaces 2017, 4, 1700507.

[37] (a) V. Postica, J. Gröttrup, R. Adelung, O. Lupan, A. K. Mishra,N. H. de Leeuw, N. Ababii, J. F. C. Carreira, J. Rodrigues,N. B. Sedrine, M. R. Correia, T. Monteiro, V. Sontea, Y. K. Mishra.Adv. Funct. Mater. 2017, 27, 1604676; (b) S. N. Das, J. P. Kar,J.-H. Choi, T. I. Lee, K.-J. Moon, J.-M. Myoung, J. Phys. Chem. C 2010,114, 1689; (c) O. Lupan, V. Postica, N. Wolff, O. Polonskyi,V. Duppel, V. Kaidas, E. Lazari, N. Ababii, F. Faupel, L. Kienle,R. Adelung, Small 2017, 13, 1602868.




Documents

Functionalized Pd/ZnO Nanowires for Nanosensorslc/2018-PSS-Pd-ZnO.pdf · 2018-01-15 · enhanced performances. A series of nanosensor devices is fabricated based on single Pd/ZnO